Proc. NatL Acad. Sci. USAVol. 79, pp. 1859-1863, March 1982Botany

Liposome-mediated delivery of tobacco mosaic virus RNA intotobacco protoplasts: A sensitive assay for monitoringliposome-protoplast interactions

(phospholipid vesicles/liposomes/intracellular delivery/nucleic acids/plant cells)

ROBERT T. FRALEY*, STEPHEN L. DELLAPORTAt, AND DEMETRIOS PAPAHADJOPOULOSt*Monsanto Company, St. Louis, Missouri 63167; tCold Spring Harbor Laboratory, Cold Spring Harbor, New York, 11724; and *Cancer Research Institute andDepartment of Pharmacology, University of California Medical Center, San Francisco, California 94143

Communicated by Howard A. Schneiderman, October 30, 1981

ABSTRACT Tobacco mosaic virus (TMV) RNA was encap-sulated in large, unilamellar phospholipid vesicles (liposomes), andthe encapsulated TMV RNA was shown to be infectious when in-cubated with tobacco protoplasts under appropriate conditions.Maximal virus production in protoplasts was observed after theirincubation with TMV RNA entrapped in phosphatidylserine/cholesterol liposomes. Infection was dependent on the presenceof polyalcohols in the incubation mixture. Other parameters, suchas the extent of vesicle binding, the cell-induced leakage of vesiclecontents, and the degree of liposome toxicity were shown to beimportant in determining the efficiency of infectivity. Liposome-mediated delivery offers an efficient and reproducible method forintroducing RNA into plant protoplasts.

Because of their ability to fuse directly with the plasma mem-brane or to be taken up by cells by an endocytosis-like process,liposomes have been used to introduce a variety of molecules,including drugs and enzymes into cultured mammalian cells(1, 2). Recent improvements in liposome technology (3, 4) nowpermit the encapsulation of large macromolecules such as nu-cleic acids in liposomes, and it has recently been shown thatliposome-encapsulated RNA and DNA molecules can be intro-duced and expressed in a variety of mammalian cell lines (forreview, see 5).

Liposomes offer several advantages as carriers for deliveringnucleic acids into cultured cells, including (i) protection of en-trapped RNA and DNA molecules from degradation by nu-cleases, (ii) low toxicity towards cells, (iii) effective utilizationwith a variety of cell types, and (iv) enhanced efficiency ofexpression of the encapsulated nucleic acid. These advantageshave been recognized by several laboratories (6-12) which haveattempted to use liposomes to introduce nucleic acids into plantprotoplasts. However, to date there have been no reports de-scribing the expression of liposome-encapsulated nucleic acidsby plant cells.We have investigated liposome-protoplast interactions in

order to optimize the delivery of liposomal contents to proto-plasts and to determine the potential of this delivery system forintroducing nucleic acids into plant cells. This study reports onoptimal conditions for delivering liposome-encapsulated to-bacco mosaic virus (TMV) RNA in a biologically-active form totobacco protoplasts as a convenient assay for monitoring thedelivery of nucleic acids into plant cells.

MATERIALS AND METHODSTMV RNA Isolation and Encapsulation in Liposomes. RNA

was extracted from virus (TMV strain U1) by the method ofDiaz-

Ruiz and Kaper (13) and stored in ethanol at -80°C until use.Liposomes were prepared by the method of Szoka and Papa-hadjopoulos (14) as modified by Fraley et al. (15) for the encap-sulation of nucleic acids. Briefly, 500 ,ug ofTMV RNA and 0.1,Ci (1 Ci = 3.7 X 1010 becquerels) of [3H]poly(A) in 0.33 mlof sterile buffer (5 mM Tris/50 mM NaCl/0.4 M mannitol/0. 1mM EDTA, pH 7.0) was added to 10 ,umol of phospholipiddissolved in diethyl ether (1.0 ml). The sources, purity, andstorage oflipids used in this study have been reported (16). Thetwo-phase solution was briefly sonicated (==5 sec) in a bath-typesonicator and transferred to a rotary evaporator, and the etherwas removed under reduced pressure as described (15). Lipo-somes used for binding studies were prepared in the same man-ner except that either [3H]dipalmitoyl phosphatidylcholine(PtdCho) (50 ,uCi) or [3H]inulin (50 MCi) was included in thepreparation as radioactive labels for the vesicle lipid or aqueousinterior, respectively.

Lipid concentrations were measured by phosphate analysis(17), and RNA encapsulation was determined by monitoring analiquot (25-50 ,ul) for radioactivity. Typically, between 30-40%of the TMV RNA was encapsulated by this procedure. The in-tegrity of the encapsulated RNA was determined by extractingthe RNA from liposomes by the method of Bligh and Dyer (ex-cept that isopropanol was substituted for methanol; ref. 18) andanalyzing it by gel electrophoresis (19).

Protoplast Isolation, Incubation, and Culture. Tobacco pro-toplasts were isolated from suspension cultures ofNicotiana ta-bacum "Xanthi" (n. c.) as described by Uchimiya and Murashige(20) and were resuspended at a final cell density of2 x 106 cellsper ml in Tris-buffered mannitol solution (TBM solution; 5mMTris/0.5 M mannitol/0.5 mM CaCl2, pH 7.0). Liposomes(10-500 nmol of phospholipid) were added to 0.5 ml of proto-plasts (in a 15.0-ml plastic centrifuge tube), mixed, and allowedto incubate for 5.0 min at 25°C. TBM (4.5 ml) containing 5%(vol/vol) glycerol, 5% (vol/vol) dimethyl-sulfoxide, 5% (vol!vol) ethylene glycol, 10% (wt/vol) polyvinyl alcohol (PVA)(Polysciences, no. 2975), or 10% (wt/vol) polyethylene glycol(PEG) (Mr = 6000; British Drug House, Poole, England) wasadded to the protoplast suspension and allowed to incubate for20 min. The viscous solutions then were diluted by the additionof 10 ml ofTBM solution, and the protoplasts were pelleted bycentrifugation (100 X g for 10 min). The protoplasts werewashed twice with TBM solution, once with MS medium [ref.21; B5 vitamin (22)/0.5 M mannitol/1.5% sucrose/2 ,ug of (2,4-

Abbreviations: TMV, tobacco mosaic virus; TBM solution, Tris-buf-fered mannitol solution; PtdCho, phosphatidyl-choline from hen eggyolks; PtdSer, phosphatidylserine from bovine brain; SteN, stearyl-amine; PVA, polyvinyl alcohol; PEG, polyethylene glycol; Chol, cho-lesterol; LD25, lipid concentration that reduces cell viability by 25%.

1859

The publication costs ofthis article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 79 (1982)

dichlorophenoxy)acetic acid per ml] and were resuspended ata final cell density of 105 cells per ml in medium. The protoplastsuspension (2.0 ml) was dispensed into tissue culture flasks(Falcon T3025) and cultured in the dark for 24 hr before transferto a light intensity of approximately 2000 lx. Aliquots (100 tulin duplicate) were removed at 48 hr and stored (-200C) fordetermination of virus levels.The determination of cell-associated vesicle lipid and vesicle

contents was as described except that after liposome/cell in-cubation, the protoplasts were washed twice by centrifugationin TBM solution containing 2% Ficoll (the liposomes cannotenter solutions of this density, whereas the protoplasts pelletduring centrifugation). The cells were resuspended in 1.0 mlof TBM solution and transferred directly to scintillation vialsfor determination of cell-associated radioactivity.

Radioimmunoassay for Monitoring TMV Production in Pro-toplasts. Rabbit anti-TMV antiserum was prepared as described(23), and iodinations were performed with Enzymo Beads (Bio-Rad) according to the manufacturer's directions. Microtiterplates (Dynatech Laboratories, Alexandria, VA) were precoatedwith unlabeled anti-TMV antibody (10 yg/ml in 50 mM phos-phate buffer, pH 9.0) and washed three times with phosphate-buffered saline (containing 1% bovine serum albumin) beforeuse. Aliquots of protoplasts (subjected twice to freeze/thawcycle) or dilutions of TMV were added to wells and incubatedfor 12-18 hr at 250C; the wells were washed three times withphosphate-buffered saline (containing 1% bovine serum albu-min), and 125 u1 of the Iff-labeled antibody (r200,000 cpm)was added and allowed to incubate for 3.0 hr. The wells were

washed five times with phosphate-buffered saline, separated,and transferred to scintillation vials for the determination ofradioactivity. The assay detects 1 ng of virus (at twice back-ground levels); at saturating amounts of virus (> 1 pg), =3-5%of the labeled antibody was bound. The reproducibility of theresults was quite good; all experimental points were determinedin triplicate, and a variation of <14% was observed betweenseparate samples. The values from independent experimentsagreed within a factor of 2.

RESULTSA variety of different liposome preparations, including neutral(6, 10), negative (8-11), and positively charged (7, 10, 12) ves-icles, have been reported to have been successfully used to in-troduce encapsulated materials into plant protoplasts. TMVRNA was encapsulated in each of these different vesicle typesto determine which liposome composition was optimal for de-livery of TMV RNA into protoplasts (Table 1). However, pre-liminary experiments in which tobacco protoplasts were simplyincubated with these various liposome preparations resulted inno detectable virus production. The lack of infectivity of theliposome-encapsulated TMV RNA was not attributable to deg-radation because analysis of the encapsulated RNA on agarosegels revealed no breakdown had occurred during encapsulation(data not shown), and infection with the free RNA could be ob-tained by using the poly-L-ornithine method. Instead, a num-ber of factors, including (i) inefficient binding of liposomes toprotoplasts, (ii) leakage of encapsulated RNA from liposomesduring incubation with cells, (iii) failure of liposomes to intro-duce their contents intracellularly, and (iv) possible toxic effectsof liposomes on cells could account for the lack of infectivity ofthe encapsulated TMV RNA. These factors were examined todetermine their possible influence on liposome-protoplastinteractions.

Association of Vesicle Lipid and Internal Contents with Pro-toplasts. Liposomes containing [3H]dipalmitoyl phosphatidyl-choline as a trace radioactive phospholipid label were incubated

Table 1. Infectivity of TMV RNA encapsulated in variousliposome preparations

Vesicle lipid Encapsulationcomposition efficiency* Infectivityt Infectivity*

PtdSer 15.2 <1 85PtdCho 17.7 <1 <3PtdCho/SteN 18.3 <1 <1PtdSer/Chol 16.4 <1 503

TMVRNA (500 pug) was encapsulated in liposomes (10 jtmol of lipid)prepared by reverse-phase evaporation. A small amount of radioactive[3Hlpoly(A) is included to permit precise calculations of encapsulationefficiency and TMV RNA concentrations. Unencapsulated TMV RNAwas separated from liposomes by centrifugation (flotation) on discon-tinuous Ficoll gradients (15). Ficoll solutions were prepared in 5 mMTRIS/0.5 M mannitol/0.1 mM EDTA, pH 7.0. Rapidly growing sus-pension cultures were centrifuged (200 x g), resuspended at twice theoriginal volume in 2% Cellulysin (Calbiochem)/1% Driselase (KyowaHakko Kogyo, Tokyo, Japan)/0.5% Macerase (Calbiochem)/0.5 Mmannitol (pH 5.7), and incubated for 5-6 hr. The resulting protoplastswere separated from debris and undigested cells by successive passagethrough 100, 150, and 200 mesh filters (Small Parts, Miami, FL) andwere washed twice by centrifugation (100 x g for 5 min) with the abovebuffer. Protoplasts (106) were incubated with 400-500 nmol of lipid(5 jtg of RNA) in the presence and absence of 10% polyvinyl alcoholand were assayed 48 hr later for virus production.*The amount of encapsulated RNA is expressed as jg/flmol ofphospholipid.

t The infectivity of liposome-encapsulated TMV RNA in the absenceof PVA treatment (ng of virus per 105 protoplasts).

t The infectivity of liposome-encapsulated TMV RNA after treatmentwith 10% PVA (ng of virus per 105 protoplasts).

with 2 x 105 protoplasts, and the amounts of cell-associatedvesicle lipid remaining after extensive washing of the proto-plasts was determined (Fig. LA). It should be emphasized thatthis experimental approach does not distinguish between li-posomes adsorbed to the cell surface and those taken up byprotoplasts by fusion, endocytosis, or some other process;therefore, these vesicles are referred to as cell-associated. Pos-itively charged liposomes [PtdCho/stearyl amine (SteN) lipo-somes] were found to associate with protoplasts to a muchgreater degree than either neutral PtdCho or negativelycharged [phosphatidylserine (PtdSer)] liposomes (Fig. LA). Thelevel of binding of the various liposome preparations (Fig. LA)was =10- to 100-fold greater than had been observed previouslywith mammalian cells (15, 24, 25), suggesting that inefficientbinding is not a likely cause for the lack of infectivity of encap-sulated TMV RNA.

Because liposomes are known to exhibit increased perme-ability during their incubation with both mammalian cells (15)and plant protoplasts (6), it is important to measure the amountof vesicle contents that become cell-associated to determine ifcell-induced leakage can account for the lack of infectivity. Forthis purpose, liposomes containing [3H]inulin (Fig. 1B) wereused to determine the association of vesicle contents with pro-toplasts under conditions identical to those used in Fig. LA todetermine vesicle lipid association with cells. The use of inulinas a marker for vesicle internal contents is preferable to usingradioactively-labeled RNA because it eliminates possible arti-facts associated with the binding or uptake of free RNA (or nu-cleotides) by protoplasts, which may have leaked from the li-posomes during the incubation. The amount of [3H]inulin(expressed as vesicle lipid equivalents) associated with proto-plasts was highest for the PtdCho/SteN liposomes (Fig. LA). Bycomparing the amounts of cell-associated vesicle lipid and con-tents, it could be determined that PtdCho/SteN liposomes re-tained 48% of their contents after incubation with protoplasts.PtdSer and PtdCho liposomes retained 10.4% and 30% of their

1860 Botany: Fraley et aL

Proc. Natl. Acad. Sci. USA 79 (1982) 1861

X 15r,-.

8 - 10"0r"0,a *H

co .,.

5U,..P"-

._4

w 0 15as

X 10

ia

v v 5

A >

I "A4laB

SAC-- A-

200 400Phospholipid, nmol

FIG. 1. Association of vesicle lipid and internal contents with pro-toplasts. Liposomes, containing either [3H]dipalmitoyl PtdCho (5 4Ci/pmol oflipid) or [3H]inulin (0.15 ACi/ul) to labelvesicle lipidorinternalcontents, respectively, were prepared. Liposomes (10-500 nmol oflipid) were incubated with protoplasts (2 x 105 cells) for 20 min. Theamount of cell-associated radioactivity was determined after centrif-ugation of protoplasts through 2% (wt/vol) Ficoll buffer. (A) Proto-plast-associated vesicle lipid (nmol) after incubation of protoplastswith increasing amounts of radioactively labeled PtdSer (0), PtdCho(r), PtdCho/SteN (A), and PtdSer/Chol (e) liposomes. (B) Protoplast-associated vesicle contents (expressed as vesicle lipid equivalents)after incubation of protoplasts with increasing amounts of radioac-tively-labeled liposomes.

contents, respectively. These results confirm and extend thoseof Lurquin (6), which demonstrated that a large fraction of li-

posome contents (-70%) are released during a short incubationwith cowpea protoplasts, and they indicate that vesicle leakageis an important parameter that could influence the efficiencyof liposomal delivery. However, the problems of leakage couldbe minimized by the addition of Chol during liposomepreparation.

Effects of Different Liposome Preparations on ProtoplastViability. Several investigators (7, 10) have reported significantloss ofprotoplast viability after their incubation with liposomes,and this could dramatically influence the apparent efficiencyof delivery as monitored by TMV RNA infectivity. The effectsof increasing liposome concentrations on protoplast viabilitywere determined by using the fluorescein diacetate technique(26) to monitor viable cells 48 hr after incubation with liposomes(Fig. 2). Positively charged liposomes were found to be more

toxic (LD25 = 135 nmol of lipid) than either neutral (LD25 =

645 nmol of lipid) or negatively charged liposomes (LD25 = 330nmol of lipid); however, at lower lipid concentrations (100-200nmol) these cytotoxic effects were minimal and would not beexpected to interfere with virus replication.

Effect of Lipid Composition and Incubation Conditions on

Liposome-Mediated Delivery of TMV RNA to Protoplasts.These results show that a significant amount of liposome con-

tents became cell-associated under conditions in which proto-plast viability was maintained, yet the lack of detectable virusproduction indicates that significant intracellular delivery oftheencapsulated TMV RNA did not occur. The efficiency of lipo-some-mediated delivery to various mammalian cell lines hasbeen shown to be increased greatly by exposing the cells toagents such as glycerol, PEG, dimethyl sulfoxide, and ethylene

o80--t 80 LD25C70-A

60

50

~04a

30

200 400Phospholipid, nmol

FIG. 2. Effect of different liposome preparations on protoplast vi-ability. Empty liposomes were incubated with protoplasts (1 x 106cells) as described in the legend to Table 1. Cells were then washed andresuspended in medium (2 x 10' cells per ml). Forty-eight hours afterthe incubation with liposomes, the percentage of viable protoplasts(expressed as % of control viability) was determined by using the flu-orescein diacetate technique (26). The average viability of nontreatedcontrol cells was -80%. o, PtdSer; n, PtdCho; A, PtdCho/SteN. LD25,the lipid concentration which reduces cell viability by 25%.

glycol after their incubation with liposomes (ref. 15, unpub-lished data). Therefore, a variety of postincubation treatmentswere examined for their ability to stimulate liposome deliveryand to increase the infectivity of encapsulated TMV RNA. Ofthe various treatments examined, exposure of tobacco proto-plasts to dilute PVA solutions proved most effective in stimu-latingTMV RNA infectivity and was studied in detail (Table 2).The PVA treatment was specific for PtdSer liposomes (Table 1),and the relative infectivities of the various liposome prepara-tions at different liposome concentrations is shown in Fig. 3.The infectivity of PtdSer liposomes could be enhanced sub-stantially (5- to 6-fold) by including an equimolar concentrationof cholesterol (Chol) in the preparation (Table 1 and Fig. 3).Chol had no effect on vesicle binding to protoplasts (Fig. 1A)but significantly increased the amount of vesicle contents thatbecame cell-associated (Fig. 1B) from 10.4% to 90%. AddingChol to either neutral or positively charged liposome prepa-rations had no effect on delivery (unpublished observations).

Table 2. Effect of various treatments on the infectivity ofliposome-encapsulated TMV RNA

Treatment Infectivity* Referencet

Glycerol <1 15PEG 205 6,15Dimethyl sulfoxide <1 15Ethylene glycol <1 Unpublished dataPVA 503 27

PtdSer/Chol liposomes containing 16.9 ug of RNA per tumol of lipidwere prepared. Vesicle lipid (400 nmol; 5 ,ug RNA) was incubated withprotoplasts (2 x 106) for 5.0 min in 0.5 ml of TBM. TBM (4.5 ml) con-taining glycerol, PEG, dimethyl sulfoxide, ethylene glycol, orPVA wasadded to the cells and allowed to incubate for an additional 20 min.Protoplasts were washed, resuspended in media, and assayed 48 hrlater for virus production.* Expressed as ng of virus per 105 protoplasts. Conditions were de-scribed in Table 1 and in Materials and Methods.

t Previous reference to the use of these treatments either with mam-malian cells (ref. 15; unpublished data) or plant protoplasts (6, 27).

4.

Botany: Fraley et aL

1[

Proc. Natl Acad. Sci. USA 79 (1982)

"~400 -

300

200 -

100

200 400Phospholipid, nmol

FIG. 3. Infectivity of TMV RNA encapsulated in various liposomepreparations. Liposomes containing TMV RNA were prepared as de-scribed in the legend to Table 1. Vesicle lipid (10-500 nmol) was in-cubated with protoplasts (1 x 106 cells) for 5.0 min prior to the additionof 4.5 ml of a PVA solution (10% wt/vol). The cells were washed, re-suspended in media, and assayed 48 hr for virus production. o, PtdSer;a, PtdCho; A, PtdCho/SteN; e, PtdSer/Chol liposomes.

It should also be noted that the PVA treatment enhanced theinfectivity of TMV but not that of TMV RNA (Table 3). Thatliposomes are directly involved in the delivery process (i.e.,fusion or endocytosis) is suggested by the observations (Table3) that empty liposomes did not promote the infectivity of freeTMV RNA and that encapsulated TMV RNA was introducedinto protoplasts by a mechanism that protects the RNA fromdigestion by ribonuclease added to the medium.

DISCUSSIONIn this paper we demonstrated the expression of liposome-en-capsulated nucleic acids in plant protoplasts. Several studieshave shown that radioactively labeled nucleic acids entrappedin liposomes can become cell-associated after their incubationwith protoplasts (6-11), and cell fractionation and autoradiog-raphy have been used in attempts to localize the labeled nucleicacids within the cell. Although the results of these earlier stud-

Table 3. Infectivity of TMV and TMV RNA under variousincubation conditions

Incuba- RNA, Infec-RNA preparation tion* lug RNaset tivityt

TMV (whole virus) - 0.25 - <1PLO 0.25 - 6000PVA 0.25 - 90

TMVRNA - 5.0 - <1PLO 5.0 - 135PVA 5.0 - <1PLO 5.0 + 5

TMV RNA plusempty PtdSer LUV PLO 5.0 - <1

TMV RNA in PtdSer/Chol LUV - 5.0 - <1

PVA 5.0 - 480PVA 5.0 + 450

Protoplast incubations and virus assay were carried out as describedin the text. Liposomes containing TMV RNA were prepared fromPtdSer and Chol and contained 16.9 .g ofTMVRNA per ,umol of lipid.Empty liposomes, prepared identically, were added at a concentrationof 450 nmol per incubation. LUV, large unilamellar vesicles; PLO,poly(L-ornithine).$ Treatment with PVA was described in the text; incubations in thepresence of PLO were as in ref. 28.

tRNA preparations were preincubated with ribonuclease (10 Ag) for15 min prior to their addition to protoplasts.

t Expressed as ng of virus per 105 protoplasts.

ies were encouraging, they also were ambiguous because theassays used could not distinguish between adsorption of lipo-somes to the cell surface and intracellular delivery (1, 2). Inaddition, it has been demonstrated that damaged protoplasts(29) and isolated nuclei (30) may bind substantial amounts ofnucleic acids. As a result, inference that the intracellular deliv-ery of radioactive DNA sequestered in liposomes has occurredbased strictly on the localization of radioactivity in nuclei couldgive misleading results. The use of a biological assay for mon-itoring liposome-mediated delivery overcomes these uncertain-ties and allows for the unambiguous determination ofconditionsthat favor intracellular delivery.Our results indicate that negatively charged (PtdSer/Chol)

liposomes are the most efficient carrier for introducing TMVRNA into tobacco protoplasts. This observation is somewhatsurprising because protoplasts are highly negatively charged(31), but the superiority of PtdSer/Chol liposomes can at leastbe partially understood in terms of a combination of their highaffinity for cells (Fig. 1A), resistance to protoplast-induced leak-age ofvesicle contents (Fig. 1B), and their relatively low toxicity(compared with positively charged liposomes) to protoplasts(Fig. 2). These observations are not limited to tobacco cells be-cause similar results have been obtained for the delivery of li-posome-encapsulated cowpea mosaic virus RNA to cowpea pro-toplasts (unpublished data).The role of Chol in enhancing the infectivity ofliposome-en-

capsulated TMV RNA correlates with its ability to reduce theextent of vesicle leakiness during incubation with cells. There-fore, the net result is an increased amount ofencapsulatedTMVRNA that remains available for intracellular delivery. Similareffects of Chol in enhancing simian virus 40 DNA delivery tomammalian cells have been observed (15). These results are inconflict with those of Ostro et al. (8), which indicate that theinclusion of Chol in negatively-charged liposomes actually de-creases the extent of liposomal delivery. It should be empha-sized, though, that the latter study utilized an assay based onthe uptake of radioactively labeled DNA.The absolute efficiency of virus production after liposome-

mediated delivery ofTMV RNA to protoplasts (-400-500 ngof virus per 105 protoplasts; Table 3) is comparable or greaterthan values reported forTMV RNA infection mediated by poly-cations (Table 2; ref. 28) or alkaline pH (32); however, absolutecomparisons are difficult to make because of the different cellpreparations used in these studies (suspension cells vs. leaf tis-sue). Under optimal conditions, =20% of the tobacco proto-plasts become stained with fluorescein-labeled anti-TMV an-tibody (33) 48 hr after incubation with liposome-encapsulatedTMV RNA. Other important advantages of liposome-mediateddelivery are its high degree of reproducibility and the fact thatonce the RNA is encapsulated in liposomes, it remains stablefor several weeks and may be used for many experiments. Fi-nally, although minor changes in incubation conditions are re-quired for maximal liposome delivery in other protoplast sys-tems (cowpea, petunia, tobacco mesophyll, etc.), this methodappears to be generally applicable to use with all plant species.The mechanism by which the PVA and PEG treatments stim-

ulate intracellular delivery ofliposomal contents is not yet clear.These agents are known to induce protoplast-protoplast fusion(27, 34) and also may be able to facilitate liposome-protoplastfusion as well. Alternatively, PEG and PVA may stimulate li-posome uptake by an endocytosis-like process, as has been sug-gested for the action ofglycerol in enhancing liposome deliverywith mammalian cells (15, 35). Although such processes havenot been clearly elucidated in plant cells, there have been sev-eral reports describing the uptake of large particles by proto-plasts (34, 36), particularly after treatment with PEG. Our ef-

1862 Botany: Fraley et aL

Proc. Natl Acad. Sci. USA 79 (1982) 1863

forts are directed towards understanding these aspects ofliposome-protoplast interaction and increasing further the ef-ficiency of liposome-mediated delivery to plant cells.

We wish to thank Drs. Eugene Nester, Milton Gordon, and KennethGiles for their help and grant support for this project and Ms. Jean Swal-low and Mr. Brian Watson for their assistance with the preparation ofthe manuscript. This work was supported by grants from the NationalCancer Institute (CA 25526), the General Medical Institute (GM28117), the National Science Foundation (PCM-800-3807), and the JaneCoffin Child's Memorial Fund for Medical Research (RTF).

1. Poste, G. (1980) in Liposomes in Biological Systems, eds. Gre-goriadis, G. & Allison, A. C. (Wiley, New York), pp. 101-151.

2. Pagano, R. & Weinstein, J. (1978) Annu. Rev. Biophys. Bioeng.7, 435-468.

3. Szoka, F. & Papahadjopoulos, D. (1980) Annu. Rev. Biophys.Bioeng. 9, 467-508.

4. Deamer, D. & Uster, P. (1980) in Introduction of Macromole-cules into Viable Mammalian Cells, eds. Baserga, R., Croce, C.& Rovera, G. (Alan R. Liss, New York), pp. 205-220.

5. Fraley, R. & Papahadjopoulos, D. (1981) Trends Biochem. Sci. 6,77-80.

6. Lurquin, P. (1979) Nucleic Acids Res. 12, 3773-3784.7. Lurquin, P. (1981) Plant Sci. Lett. 21, 31-40.8. Ostro, M., Lavelle, D., Paxton, W., Matthew, B. & Giacomoni,

D. (1980) Arch. Biochem. Biophys. 201, 392-402.9. Matthew, B., Dray, S., Widholm, J. & Ostro, M. (1979) Planta

145, 37-44.10. Rollo, F., Sala, F., Cella, R. & Parisi, B. (1980) in Plant Cell Cul-

tures, Results and Perspectives, eds. Sala, F., Parisi, B., Cella,R. & Ciferri, 0. (Elsevier-North Holland, New York), pp.237-246.

11. Rollo, F., Galli, M. & Parisi, B. (1981) Plant Sci. Lett. 20,347-354.

12. Casells, A. (1978) Nature (London) 275, 760.13. Diaz-Ruiz, J. & Kaper, J. (1978) Prep. Biochem. 8, 1-17.

14. Szoka, F. & Papahadjopoulos, D. (1978) Proc. Natl Acad. Sci.USA 75, 4194-4198.

15. Fraley, R., Subramani, S., Berg, P. & Papahadjopoulos, D.(1980) J. Biol. Chem. 255, 10431-10435.

16. Mayhew, E., Rustum, Y., Szoka, F. & Papahadjopoulos, D.(1979) Cancer Treat. Rep. 63, 1923-1928.

17. Bartlett, G. (1959) J. Biol. Chem. 234, 466-468.18. Bligh, E. & Dyer, W. (1959) Can. J. Biochem. Physiol 37,

911-917.19. Locker, J. (1979) AnaL Biochem. 98, 358-367.20. Uchimiya, H. & Murashige, T. (1974) Physiol, Plant 54, 936-943.21. Murashige, T. & Skoog, F. (1962) Physiot Plant 15, 473-497.22. Gamborg, O., Miller, R. & Jima, K. (1968) Exp. Cell Res. 50,

151-158.23. Matthews, R. (1967) Methods Virol, eds. Maramorosch, K. &

Koprowski, H. (Academic, New York), Vol. 3, pp. 199-241.24. Szoka, F., Jacobson, K. & Papahadjopoulos, D. (1979) Biochim.

Biophys. Acta 551, 295-303.25. Szoka, F., Jacobson, K., Derzko, Z. & Papahadjopoulos, D.

(1980) Biochim. Biophys. Acta 600, 1-18.26. Widholm, J. (1972) Stain Technol 47, 189-194.27. Nagata, T. (1978) Z. Naturwiss. Med. Grundlagenforsch. 65,

263-264.28. Takebe, I. (1977) in Comprehensive Virology, eds. Fraenkel-

Conrat, H. & Wagner, R. (Plenum, New York), Vol. 2, pp.237-283.

29. Hughes, B., White, F. & Smith, M. (1978) Plant Sci. Lett. 11,199-206.

30. Ohyama, K., Pelcher, L. & Horn, D. (1977) Plant PhysioL 60,98-104.

31. Nagata, T. & Melchers, G. (1978) Planta 142, 235-238.32. Sarkar, S., Upadhya, M. & Melchers, G. (1974) Mol Gen. Genet.

135, 1-9.33. Otsuki, Y. & Takebe, I. (1969) Virology 38, 497-501.34. Suzuki, M., Takebe, I., Kajita, S., Honda, Y. & Matsui, C. (1977)

Exp. Cell Res. 105, 127-135.35. Fraley, R., Straubinger, R., Rule, G., Springer, L. & Papahad-

jopoulos, D. (1981) Biochemistry 20, 6978-6987.36. Cocking, E. (1970) lnt. Rev. Cytol 28, 89-112.

Botany: Fraley et d